Business data processing model training method

By dynamically adjusting the weights and selection of multiple candidate training algorithms in the machine learning solver and optimizing the training process using the Nesterov acceleration algorithm, the problem of inaccurate selection of machine learning algorithms is solved, and efficient training of business data processing models is achieved.

CN118940855APending Publication Date: 2024-11-12SHANSHU TECH (BEIJING) CO LTD +5
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410920538.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-10
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

The selection process for existing machine learning algorithms relies on human judgment, resulting in insufficient robustness of algorithm performance and difficulty in ensuring the efficiency and accuracy of the training process.

Method used

By setting multiple candidate training algorithms in the machine learning solver, the top K algorithms are extracted from the largest to the smallest weight values ​​for solving, and the algorithm with the smallest error is selected as the target training algorithm. The algorithm weights are adjusted during the training process to optimize the algorithm selection, and the Nesterov acceleration algorithm is used to accelerate the training.

Benefits of technology

It improves the accuracy of machine learning algorithm selection and the efficiency of the training process, ensuring that efficient algorithms are matched to different training stages, and enhances the training efficiency of business data processing models.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118940855A_ABST
    Figure CN118940855A_ABST
Patent Text Reader

Abstract

The invention discloses a business data processing model training method, and the method comprises the steps: extracting the first K to-be-selected training algorithms from large to small according to the weight values, used for representing the solving efficiency of the to-be-selected training algorithms, of N to-be-selected training algorithms in a machine learning solver; respectively solving sub-problems corresponding to the business data processing model by using K to-be-selected training algorithms; executing a training task on the business data processing model based on the target training algorithm with the minimum solving error; and after the training task is executed for a first preset duration, correspondingly adjusting the weight value of each to-be-selected training algorithm according to the solving error of each to-be-selected training algorithm, and returning to execute the step of extracting the first K to-be-selected training algorithms from large to small until a preset condition is met, thereby obtaining a trained business data processing model. Through the technical scheme provided by the invention, the algorithm selection accuracy of machine learning can be improved, and it is ensured that high-efficiency algorithms can be matched in different training stages.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application belongs to the field of computer and data processing technology, and in particular, relates to a business data processing model training method, device, program product, medium and electronic equipment. Background Art

[0002] In recent years, machine learning has made great progress in various fields, and more and more companies are applying machine learning to conduct business prediction and analysis. Existing mainstream machine learning solvers such as H2O / Scikit learn / TensorFlow are generally based on a specific training algorithm, such as (stochastic) gradient method, conjugate gradient method, ADMM, quasi-Newton method, etc. The selection of these algorithms requires users to make their own judgment and selection, and the judgment and selection process of the algorithm requires rich experience.

[0003] However, manual selection can only design simple experiments to estimate the quality of the algorithm, but the estimation results are not accurate enough, and the algorithm performance is not robust enough. An algorithm that is very efficient at the beginning may become less efficient later. Based on this, how to improve the accuracy of algorithm selection in machine learning is a technical problem that needs to be solved urgently. Summary of the invention

[0004] The embodiments of the present application provide a business data processing model training method, device, computer program product, computer-readable storage medium and electronic device, which can improve the accuracy of machine learning algorithm selection to at least a certain extent, and ensure that high-efficiency algorithms can be matched at different training stages.

[0005] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by the practice of the present application.

[0006] According to a first aspect of an embodiment of the present application, a business data processing model training method is provided, the method comprising: inputting a business data processing model to be trained into a machine learning solver, the machine learning solver comprising N candidate training algorithms, the training algorithms being used to train the business data processing model, N being an integer greater than or equal to 2; extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms, wherein K is an integer greater than or equal to 1, K is less than N, and the weight value is used to characterize the solution efficiency of the candidate training algorithm; using the K candidate training algorithms to respectively train the business data processing model The corresponding sub-problems are solved, and the solution errors of each candidate training algorithm are obtained; the training algorithm with the smallest solution error is determined from the K candidate training algorithms as the target training algorithm, and the training task of the business data processing model is performed based on the target training algorithm; after the training task is executed for a first preset time, the weight value of each candidate training algorithm is adjusted according to the solution error of each candidate training algorithm, and the step of extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms is returned to execute, until the preset conditions are met, and the trained business data processing model is obtained.

[0007] In some embodiments of the present application, based on the aforementioned scheme, before extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms for the first time, the method also includes: randomly assigning a weight value to each of the N candidate training algorithms.

[0008] In some embodiments of the present application, based on the aforementioned scheme, the use of the K selected training algorithms to respectively solve the sub-problems corresponding to the business data processing model includes: using the K selected training algorithms to respectively solve the sub-problems corresponding to the business data processing model within the same second preset time length.

[0009] In some embodiments of the present application, based on the aforementioned scheme, the weight value of each candidate training algorithm is adjusted according to the solution error of each candidate training algorithm, including: if the solution error is less than the preset error value, the weight value is increased; if the solution error is greater than the preset error value, the weight value is lowered; if the solution error is equal to the preset error value, the weight value is not adjusted.

[0010] In some embodiments of the present application, based on the aforementioned scheme, the machine learning solver also includes a preset acceleration algorithm, which is used to accelerate the training of the business data processing model by the target training algorithm.

[0011] In some embodiments of the present application, based on the aforementioned solution, the acceleration algorithm is the Nesterov acceleration algorithm.

[0012] According to a second aspect of an embodiment of the present application, a business data processing model training device is provided, the device comprising: an input unit, used to input the business data processing model to be trained into a machine learning solver, the machine learning solver comprising N candidate training algorithms, the training algorithm being used to train the business data processing model, N being an integer greater than or equal to 2; an extraction unit, used to extract the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms, wherein K is an integer greater than or equal to 1, K is less than N, and the weight value is used to characterize the solution efficiency of the candidate training algorithm; a solution unit, used to use the K candidate training algorithms to respectively train the business data The sub-problems corresponding to the processing model are solved, and the solution errors of each candidate training algorithm are obtained; an execution unit is used to determine the training algorithm with the smallest solution error from the K candidate training algorithms as the target training algorithm, and execute the training task of the business data processing model based on the target training algorithm; an adjustment unit is used to adjust the weight values ​​of each candidate training algorithm according to the solution errors of each candidate training algorithm after the training task is executed for a first preset time period, and return to execute the step of extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms, until the preset conditions are met, and the trained business data processing model is obtained.

[0013] According to a third aspect of an embodiment of the present application, a computer program product is provided, which includes computer instructions, which are stored in a computer-readable storage medium and are suitable for being read and executed by a processor, so that a computer device having the processor executes to implement the operations performed by the method described in any one of the first aspects above.

[0014] According to a fourth aspect of an embodiment of the present application, a computer-readable storage medium is provided, in which at least one computer program instruction is stored. The at least one computer program instruction is loaded and executed by a processor to implement the operations performed by the method described in any one of the first aspects above.

[0015] According to a fifth aspect of an embodiment of the present application, an electronic device is provided, comprising one or more processors and one or more memories, wherein at least one computer program instruction is stored in the one or more memories, and the at least one computer program instruction is loaded and executed by the one or more processors to implement the operations performed by the method described in any one of the first aspects above.

[0016] Based on the technical solution proposed in this application, in the machine learning solver, according to the weight values ​​of the N training algorithms to be selected, which are used to characterize the solution efficiency of the training algorithms to be selected, the first K training algorithms to be selected are extracted from large to small, and the sub-problems corresponding to the business data processing model are solved respectively by using the K training algorithms to be selected, and the target training algorithm with the smallest solution error is determined from the K training algorithms to be selected, and the training task of the business data processing model is performed based on the target training algorithm. After the training task is executed for a first preset time, the weight values ​​of each training algorithm to be selected are adjusted according to the solution error of each training algorithm to be selected, and then the above training steps are repeated. In this way, the training algorithms to be selected are continuously adjusted during the business data processing process, the accuracy of the algorithm selection of machine learning is improved, and it is ensured that different training stages can match high-efficiency algorithms, ensuring a high training efficiency of the business data processing model.

[0017] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The drawings herein are incorporated into the specification and constitute a part of the specification, showing embodiments consistent with the present application, and together with the specification, are used to explain the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0019] Figure 1 A flowchart of a business data processing model training method in an embodiment of the present application is shown;

[0020] Figure 2 A block diagram of a business data processing model training device in an embodiment of the present application is shown;

[0021] Figure 3 A schematic diagram of the structure of an electronic device in an embodiment of the present application is shown. DETAILED DESCRIPTION

[0022] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0023] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner. In the following description, many specific details are provided to provide a full understanding of the embodiments of the present application. However, those skilled in the art will appreciate that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps, etc. can be adopted. In other cases, known methods, devices, realizations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0024] The block diagrams shown in the accompanying drawings are only functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software form, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices. It should also be noted that in the accompanying drawings, in order to ensure the simplicity of the drawings, some components in the drawings that do not affect the explanation of the technical solution of the present application are adaptively omitted.

[0025] The flowcharts shown in the accompanying drawings are only exemplary and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps can be decomposed, and some operations / steps can be combined or partially combined, so the actual execution order may change according to actual conditions.

[0026] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the feature. In the description of this application, unless otherwise specified, "plurality" means two or more.

[0027] Reference Figure 1 , which shows a flow chart of a business data processing model training method in an embodiment of the present application, and the business data processing model training method can be executed by a device having a computing and processing function, referring to Figure 1 As shown, the business data processing model training method includes at least steps S10 to S50, which are described in detail as follows:

[0028] Step S10, inputting the business data processing model to be trained into a machine learning solver, wherein the machine learning solver includes N training algorithms to be selected, and the training algorithms are used to train the business data processing model, where N is an integer greater than or equal to 2.

[0029] Step S20, according to the weight values ​​of the N candidate training algorithms, extract the first K candidate training algorithms from large to small, where K is an integer greater than or equal to 1, K is less than N, and the weight value is used to characterize the solution efficiency of the candidate training algorithm.

[0030] Step S30, using the K candidate training algorithms to solve the sub-problems corresponding to the business data processing model respectively, and obtaining the solution error of each candidate training algorithm.

[0031] Step S40, determining a training algorithm with the smallest solution error from the K candidate training algorithms as a target training algorithm, and performing a training task on the business data processing model based on the target training algorithm.

[0032] Step S50, after the training task is executed for the first preset time, the weight values ​​of each candidate training algorithm are adjusted accordingly according to the solution errors of each candidate training algorithm, and the step of extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms is returned to execute, until the preset conditions are met, and the trained business data processing model is obtained.

[0033] In the present application, the business data processing model training method can be applied to a machine learning solver, in which at least a multi-level algorithm framework can be preset. In the inner layer, the business data processing model training method proposed in the present application can be used to train the business data processing model based on the training algorithm to improve the computing efficiency. In the outer layer, an iterative calculation acceleration algorithm can be used, and the acceleration algorithm is used to accelerate the training of the business data processing model by the training algorithm. Specifically, in the present application, the acceleration algorithm can use the Nesterov acceleration algorithm.

[0034] In the present application, before extracting the training algorithms to be selected, all the training algorithms that can be used in machine learning should be added to a pre-built algorithm pool as the training algorithms to be selected. Specifically, in the present embodiment, N training algorithms to be selected can be set, and N should be greater than or equal to 2. At this time, before the above step S20, a weight can be randomly assigned to each training algorithm to be selected in the algorithm pool, or a higher weight can be assigned to an algorithm with higher efficiency based on experience.

[0035] In the present application, since the weight value is used to characterize the solution efficiency of the candidate training algorithm, in the above step S20, when extracting the candidate training algorithm, extracting the top K candidate training algorithms with corresponding weights from large to small will be conducive to selecting the candidate training algorithm with the lowest or lower solution error.

[0036] In the present application, the K selected training algorithms are used to solve the sub-problems corresponding to the business data processing model respectively, which can be to use the K selected training algorithms to solve the sub-problems corresponding to the business data processing model respectively within the same second preset time. Among them, the smaller the solution error of solving the sub-problem, the more accurate the solution result is, and it can be considered that the corresponding selected training algorithm has a higher solution efficiency.

[0037] Furthermore, the K candidate training algorithms are used to solve the sub-problems corresponding to the business data processing model within the same second preset time, and the solution results (such as solution errors) of each candidate training algorithm are obtained respectively. Among the solution errors corresponding to the K candidate training algorithms, the candidate training algorithm with the highest corresponding solution efficiency is determined. When the K candidate training algorithms are used to solve the sub-problems, they can be independently calculated in parallel. Since the solution time is the same, the size of the K solution errors can be directly compared to determine the solution efficiency of the corresponding candidate training algorithm.

[0038] In this application, by solving sub-problems, the efficiency of different candidate training algorithms in the algorithm pool is verified, and finally the training algorithm with the smallest solution error (i.e., the highest solution efficiency) is selected as the target training algorithm to ensure the subsequent efficient learning process of the business data processing model.

[0039] The inventor of the present application considers that in the entire machine learning training process, the same candidate training algorithm may produce different solution efficiencies when solving sub-problems at different training stages, that is, different training efficiencies will appear when training the business data processing model, which may cause the business data processing model to always be obtained from a small range of training algorithms. In the long run, some training algorithms with good effects but never selected may be missed, so it is necessary to explore other rarely accessed candidate training algorithms every certain period. Specifically, a cycle threshold can be set in the present application, that is, the first preset duration in step S50. When the duration threshold is exceeded, an attempt can be made to re-extract a training algorithm from the N candidate training algorithms to train the business data processing model.

[0040] In this way, the above steps S20 to S40 can be executed cyclically every preset time threshold to ensure automatic updating of the target training algorithm and ensure that the selected training algorithms used in the entire machine learning training process have high solution efficiency (i.e. training efficiency).

[0041] In the present application, in order to ensure that the next time the candidate training algorithm is extracted, the candidate training algorithm with low solution efficiency that has been verified last time will not be extracted. Before re-executing the above steps S20 to S40, a weight adjustment step can be added, namely step S50: adjusting the weight value of each candidate training algorithm according to the solution error of each candidate training algorithm.

[0042] In the present application, the weight values ​​of the training algorithms to be selected are adjusted according to the solution errors of the training algorithms to be selected, which can be specifically performed according to the following steps S51 to S53:

[0043] Step S51: if the solution error is less than a preset error value, increase the weight value.

[0044] Step S52: if the solution error is greater than a preset error value, lower the weight value.

[0045] Step S53: if the solution error is equal to the preset error value, the weight value is not adjusted.

[0046] Specifically, the weight value corresponding to the candidate training algorithm whose solution error exceeds the preset error value is reduced; and / or the weight corresponding to the candidate training algorithm whose solution error is less than the preset error value is increased. There is no restriction on the function form of weight increase and decrease. For example, a fixed value can be adjusted according to the solution error, such as the adjustment value is η. The adjustment value can also be flexibly determined according to the solution error, such as η=f 1 (z), f 1 It is an inversely correlated function of the solution error z, that is, the larger the optimization target value (the worse the result), the smaller η is, and vice versa.

[0047] In the same training process, by setting a preset error value to increase or decrease the weight of the candidate training algorithm, the candidate training algorithm in the algorithm pool can be continuously optimized. When the candidate training algorithm is randomly extracted in the subsequent process, the probability of extracting a high-efficiency algorithm can be increased, and the probability of extracting an inefficient algorithm can be reduced, thereby improving the overall computational efficiency of the machine learning process.

[0048] In order to make the embodiments of the present application easier to understand, the following is described and explained in the form of specific examples:

[0049] After each cycle completes the training of the business data processing model, the following formula can be used to evaluate the business data processing model after each training:

[0050]

[0051] Among them, l is the loss function, x is the machine learning model parameter, and ai is the sample feature, b i is the sample label, and R(x) is a penalty function of the model complexity.

[0052] The business data processing model is evaluated to obtain an evaluation result, and further, whether a preset condition is met can be judged based on the evaluation result. If the preset condition is met, the trained business data processing model is obtained.

[0053] Specifically, the method for evaluating the initial model that has completed learning with this feature can be: judging Whether the value of meets the preset convergence requirements, for example, during the convergence process, the current model parameters are adjusted to make the above If the value of is less than the preset convergence threshold, it is judged that the preset conditions are met and the current model parameters can be used as the final parameters.

[0054] In addition, in the above step S50 of the present application, in order to prevent the solver from performing optimization and feature learning without limit, a total time threshold for feature learning can be set. When the preset total time threshold is reached, the preset conditions are met, and the feature learning can also be terminated to obtain a trained business data processing model.

[0055] In addition, in the above-mentioned step S50 of the present application, the preset condition may be that when the number of cycles of the above-mentioned steps S20 to S40 reaches a preset number, a trained business data processing model is obtained.

[0056] During the feature learning process of the initial model in the solver, when the preset convergence requirements are met, the total feature learning time reaches the total time threshold, and the number of cycles reaches any one of the preset numbers, the feature learning process of the model can be terminated to obtain a trained business data processing model.

[0057] In the business scenarios of sales volume or price prediction, regression model can be used as business data processing model. Or R(x)=λ||x|| 1 .

[0058] Set up a training algorithm pool, which contains several (N) training algorithms to be selected: A 1 , A 2 , A 3 , ..., A n , the weights are set as: w 1 , w 2 , w 3 , ..., w n .

[0059] Introduce the initial variable y 0 and auxiliary variable x0 , Z 0 ; Set t=0.

[0060] Cycle steps:

[0061] Update the auxiliary variable x according to the interpolation formula t+1 =(1-α t )x t +α t Z t ;

[0062] Solve the sub-problem f determined according to the business data processing model t , (sub-problems have different variants depending on the specific problem form);

[0063] Select K algorithms with higher weights from the pool of candidate training algorithms (first target algorithms): Pair Problem f t Solve independently and in parallel, and change the weight value corresponding to the selected training algorithm according to the solution error Select the best solution There are many strategies for changing the weight value. For example, according to the multiplicative update principle, the candidate training algorithm with good performance is updated as follows: w←w·(1+η), η∈(0,1); the candidate training algorithm with poor performance is updated as follows: w←w·(1-η), η∈(0,1).

[0064] After each training task is executed for the first preset time, the first K candidate training algorithms are extracted from the largest to the smallest according to the weight values ​​of the N candidate training algorithms, and the sub-problems corresponding to the business data processing model are solved using the K candidate training algorithms respectively, and the target training algorithm with the smallest solution error is determined from the K candidate training algorithms, and the training task of the business data processing model is executed based on the target training algorithm. In this way, the candidate training algorithms are continuously adjusted during the business data processing process, ensuring a high computing efficiency.

[0065] In this application, when the preset conditions are met (the number of cycles from step S20 to step S40 above reaches the preset number or the business data processing model obtained by training is good enough), the training is terminated to obtain the trained business data processing model, otherwise the next cycle is performed: t = t + 1; finally, y is returned. t+1 .

[0066] It should be noted that the business data processing model training method proposed in this application can be applied to various business scenarios, for example, training business data processing models in business scenarios such as sales forecasting, product pricing, and store location selection.

[0067] Based on the technical solution proposed in this application, in the machine learning solver, according to the weight values ​​of the N training algorithms to be selected, which are used to characterize the solution efficiency of the training algorithms to be selected, the first K training algorithms to be selected are extracted from large to small, and the sub-problems corresponding to the business data processing model are solved respectively by using the K training algorithms to be selected, and the target training algorithm with the smallest solution error is determined from the K training algorithms to be selected, and the training task of the business data processing model is performed based on the target training algorithm. After the training task is executed for a first preset time, the weight values ​​of each training algorithm to be selected are adjusted according to the solution error of each training algorithm to be selected, and then the above training steps are repeated. In this way, the training algorithms to be selected are continuously adjusted during the business data processing process, the accuracy of the algorithm selection of machine learning is improved, and it is ensured that different training stages can match high-efficiency algorithms, ensuring a high training efficiency of the business data processing model.

[0068] The following describes an embodiment of the device of the present application, which can be used to execute the business data processing model training method in the above embodiment of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the business data processing model training method in the above embodiment of the present application.

[0069] See also Figure 2 , shows a block diagram of a business data processing model training device in an embodiment of the present application.

[0070] like Figure 2 As shown, according to the business data processing model training device 200 of the embodiment of the present application, it includes: an input unit 201, an extraction unit 202, a solution unit 203, an execution unit 204 and an adjustment unit 205.

[0071] Among them, the input unit 201 is used to input the business data processing model to be trained into the machine learning solver, and the machine learning solver includes N candidate training algorithms, and the training algorithm is used to train the business data processing model, and N is an integer greater than or equal to 2; the extraction unit 202 is used to extract the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms, wherein K is an integer greater than or equal to 1, K is less than N, and the weight value is used to characterize the solution efficiency of the candidate training algorithm; the solution unit 203 is used to use the K candidate training algorithms to solve the sub-problems corresponding to the business data processing model respectively, and Obtain the solution error of each candidate training algorithm; an execution unit 204 is used to determine the training algorithm with the smallest solution error from the K candidate training algorithms as the target training algorithm, and perform the training task of the business data processing model based on the target training algorithm; an adjustment unit 205 is used to adjust the weight value of each candidate training algorithm according to the solution error of each candidate training algorithm after the training task is executed for a first preset time, and return to execute the step of extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms, until the preset conditions are met, and the trained business data processing model is obtained

[0072] In some embodiments of the present application, based on the aforementioned scheme, the device also includes: an allocation unit, which is used to randomly assign a weight value to each of the N candidate training algorithms before extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms for the first time.

[0073] In some embodiments of the present application, based on the aforementioned solution, the solving unit 203 is configured to: use the K selected training algorithms to solve the sub-problems corresponding to the business data processing model within the same second preset time length.

[0074] In some embodiments of the present application, based on the aforementioned scheme, the adjustment unit 205 is configured as follows: if the solution error is less than a preset error value, the weight value is increased; if the solution error is greater than the preset error value, the weight value is lowered; if the solution error is equal to the preset error value, the weight value is not adjusted.

[0075] In some embodiments of the present application, based on the aforementioned scheme, the machine learning solver also includes a preset acceleration algorithm, which is used to accelerate the training of the business data processing model by the target training algorithm.

[0076] In some embodiments of the present application, based on the aforementioned solution, the acceleration algorithm is the Nesterov acceleration algorithm.

[0077] Based on the same inventive concept, an embodiment of the present application provides a computer program product, which includes computer instructions, which are stored in a computer-readable storage medium and are suitable for being read and executed by a processor, so that a computer device with the processor executes to implement the operations performed by the business data processing model training method as described above.

[0078] Based on the same inventive concept, an embodiment of the present application provides a computer-readable storage medium, in which at least one computer program instruction is stored, and the at least one computer program instruction is loaded and executed by a processor to implement the operations performed by the business data processing model training method as described above.

[0079] Based on the same inventive concept, the present application embodiment further provides an electronic device, referring to Figure 3 , shows a schematic diagram of the structure of an electronic device in an embodiment of the present application, wherein the electronic device includes one or more memories 304, one or more processors 302, and at least one computer program (computer program instruction) stored in the memory 304 and executable on the processor 302, and when the processor 302 executes the computer program, the business data processing model training method as described above is implemented.

[0080] Among them, Figure 3 In the embodiment of the present invention, a bus architecture (represented by bus 300) is shown, which may include any number of interconnected buses and bridges, and bus 300 links various circuits including one or more processors represented by processor 302 and memory represented by memory 304. Bus 300 may also link various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. Bus interface 305 provides an interface between bus 300 and receiver 301 and transmitter 303. Receiver 301 and transmitter 303 may be the same element, namely a transceiver, providing a unit for communicating with various other devices over a transmission medium. Processor 302 is responsible for managing bus 300 and general processing, while memory 304 may be used to store data used by processor 302 when performing operations.

[0081] The functions described herein may be implemented in hardware, software executed by a processor, firmware, or any combination thereof. If implemented in software executed by a processor, the functions may be stored as one or more instructions or codes on a computer-readable medium or transmitted via a computer-readable medium. Other examples and implementations are within the scope and spirit of the present application and the appended claims. For example, due to the nature of software, the functions described above may be implemented using software executed by a processor, hardware, firmware, hard wiring, or a combination of any of these. In addition, each functional unit may be integrated into a processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit.

[0082] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0083] The units described as separate components may or may not be physically separated, and the components of the control device may or may not be physical units, that is, they may be located in one place or distributed in multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0084] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly or all or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk, etc., which can store computer program instructions.

[0085] The above description is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent substitution, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of the claims of the present application.

Claims

1. A business data processing model training method, characterized in that: The method comprises: Inputting the business data processing model to be trained into a machine learning solver, wherein the machine learning solver includes N training algorithms to be selected, and the training algorithms are used to train the business data processing model, where N is an integer greater than or equal to 2; According to the weight values ​​of the N candidate training algorithms, the first K candidate training algorithms are extracted from large to small, where K is an integer greater than or equal to 1, K is less than N, and the weight value is used to characterize the solution efficiency of the candidate training algorithm; Using the K candidate training algorithms to solve the sub-problems corresponding to the business data processing model respectively, and obtaining the solution error of each candidate training algorithm; Determine a training algorithm with the smallest solution error from the K candidate training algorithms as a target training algorithm, and perform a training task on the business data processing model based on the target training algorithm; After the training task is executed for a first preset time, the weight values ​​of each candidate training algorithm are adjusted accordingly according to the solution errors of each candidate training algorithm, and the step of extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms is returned to execute until the preset conditions are met to obtain a trained business data processing model.

2. The method according to claim 1, characterized in that Before extracting the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms for the first time, the method further includes: A weight value is randomly assigned to each of the N candidate training algorithms.

3. The method according to claim 1, characterized in that The using the K selected training algorithms to respectively solve the sub-problems corresponding to the business data processing model includes: The K selected training algorithms are used to solve the sub-problems corresponding to the business data processing model within the same second preset time period.

4. The method according to claim 3, characterized in that: The step of adjusting the weight values ​​of the training algorithms to be selected according to the solution errors of the training algorithms to be selected includes: If the solution error is less than a preset error value, increase the weight value; If the solution error is greater than a preset error value, lowering the weight value; If the solution error is equal to the preset error value, the weight value is not adjusted.

5. The method according to any one of claims 1 to 4, characterized in that: The machine learning solver also includes a preset acceleration algorithm, which is used to accelerate the training of the business data processing model by the target training algorithm.

6. The method according to claim 5, characterized in that The acceleration algorithm is the Nesterov acceleration algorithm.

7. A business data processing model training device, characterized in that: The device comprises: An input unit, used to input the business data processing model to be trained into a machine learning solver, wherein the machine learning solver includes N training algorithms to be selected, and the training algorithms are used to train the business data processing model, where N is an integer greater than or equal to 2; An extraction unit, used to extract the first K candidate training algorithms from large to small according to the weight values ​​of the N candidate training algorithms, wherein K is an integer greater than or equal to 1, K is less than N, and the weight value is used to characterize the solution efficiency of the candidate training algorithm; A solving unit, used to solve the sub-problems corresponding to the business data processing model respectively by using the K candidate training algorithms, and obtain the solution error of each candidate training algorithm; An execution unit, used to determine a training algorithm with the smallest solution error from the K candidate training algorithms as a target training algorithm, and execute a training task for the business data processing model based on the target training algorithm; The adjustment unit is used to adjust the weight values ​​of each candidate training algorithm according to the solution error of each candidate training algorithm after the training task is executed for a first preset time, and return to execute the step of extracting the first K candidate training algorithms from large to small according to the weight values ​​of N candidate training algorithms, until the preset conditions are met, and the trained business data processing model is obtained.

8. A computer program product, characterized in that The computer program product includes computer instructions, which are stored in a computer-readable storage medium and are suitable for being read and executed by a processor, so that a computer device having the processor executes the method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one program code, and the at least one program code is loaded and executed by a processor to implement the operations performed by the method according to any one of claims 1 to 6.

10. An electronic device, characterized in that: The electronic device includes one or more processors and one or more memories, wherein the one or more memories store at least one program code, and the at least one program code is loaded and executed by the one or more processors to implement the method according to any one of claims 1 to 6.